Intellectual Property – Patents – Claims limited by function or result found valid.
Justice Nicholas’ last judicial act as a judge of the Federal Court of Australia was the dismissal of Sanofi’s appeal from a decision of a Delegate of the Commissioner of Patent who had rejected Sanofi’s opposition to Amgen’s five patent applications for cholesterol-lowering antibody technology. This case is part of the ongoing patent dispute between Sanofi and Amgen over the Repatha patents, which has seen parallel proceedings in other jurisdictions, including the United States and the European Union.
This case may set a precedent for claims that solely define by function or result (here, an antibody inhibiting binding of a protein by blocking or reducing binding at a surface region of that protein), and not by structure (e.g. the amino acid sequence of an antibody). However, such precedent may have restricted utility given the decision was made under the Patents Act 1990 (Cth) as it stood prior to the Intellectual Property Laws Amendment (Raising the Bar) Act 2012 (the RTB Act).
Summary of findings
Sanofi argued failure to define a manner of manufacture, lack of fair basis, lack of clarity, failure to fully describe the invention and lack of inventive step.
Justice Nicholas’ findings were:
- (definition) The claims, although not specifying the amino acid sequences of the antibodies, were sufficiently clear in defining the antibodies’ functional characteristics.
- (manner of manufacture) The claims were based on chemical compounds defined by their binding characteristic and function, not abstract information or a “mere desideratum”.
- (fair basis) The claims were fairly based on the specification which supported a broader class of antibodies beyond the specific examples disclosed.
- (clarity) The absence of a numerical threshold to define “competes”, or a practical standard to assess whether an antibody competes, did not imply a lack of clarity. Antibodies either compete or they do not, and the methods for determining competition were well-known and available at the priority date.
- (sufficiency) The specification did not need to define the antibodies by their amino acid sequences; a skilled person could still determine whether an antibody was within the claims.
- (inventive step) The claimed antibodies were not obvious to a person skilled in the art at the priority date. Despite the known role of PCSK9 in cholesterol regulation, the development of antibodies was not routine or directly suggested by the prior art.
Scientific background
Amgen’s patent applications concerned antibody technology for treating and diagnosing cholesterol-related disorders such as hypercholesterolaemia.
The antibodies in the claim inhibit PCSK9, a protein that removes LDLR from the liver’s surface. LDLR removes LDL-type cholesterol from the blood stream. When inhibited, PCSK9 prevents LDLR from removing cholesterol. Hence, the antibodies reduce cholesterol.
By way of background to the matters discussed below:
- Antibodies are proteins. Proteins are made up of amino acids. A string of amino acids is called a peptide and each amino acid in that peptide is called an amino acid residue.
- PCSK9 has regions on its surface called epitopes that antibodies bind to.
- Antibodies operate on PCSK9’s surface, unlike other proteins and molecules that operate internally.
The Patent Applications
Each relevant claim had two parts; a statement that the antibody inhibits binding of PCSK9 to LDLR; and a description of how it does so. The claims were classified into 3 categories based on the second component:
- The Epitope Claims and Residue Claims described antibodies that inhibit PCSK9 binding to LDLR by blocking or reducing binding at epitopes and residues of PCSK9, respectively.
- The Competition Claims described antibodies that compete for binding to PCSK9 with other identified antibodies.
Defining the invention
Before its amendment by the RTB Act, section 40(2)(b) of the Act required standard patents to contain claims “defining the invention”.
Sanofi argued that the specifications lacked sufficient definition of the invention, because they:
- did not provide the amino acid sequences of the antibodies; and
- would not allow a person to know, in advance of creating an antibody, whether their antibody would infringe, giving rise to the practical difficulty of not knowing whether their method of making antibodies would fall within the claims.
Nicholas J rejected these arguments, holding that:
- Defining the physical characteristics of the antibodies was impractical. The only practical approach was to describe their functional outcomes – the amino acid sequences were not essential.
- It was enough that the claims outlined the antibodies’ features and suggested methods for establishing if another antibody meets them (albeit by detailed experimentation).
- A definition is not insufficient just because deciding if there’s been an infringement is difficult.
Manner of manufacture
Sanofi argued that the invention was not a manner of manufacture because it was a “mere desideratum” or an “abstract idea” of an antibody with the desired result or outcome. It also argued that granting the patent would result in an impermissible “chilling effect” on antibody and PCSK9 research.
“Mere desideratum”
In rejecting Sanofi’s argument his Honour emphasized the practical distinction between discovery and invention. His Honour referred to two decisions in which the “mere desideratum” argument was advanced and rejected:
- In Eli Lilly and Co v Pfizer Overseas Pharmaceuticals (2005) 218 ALR 408 the claimed invention was a chemical composition that allowed oral treatment of erectile dysfunction. The court found that the invention was not a mere desideratum because it went further than method (oral treatment) and described an orally effective chemical composition.
- In Albany Molecular Research Inc v Alphapharm Pty Ltd (2011) 90 IPR 457 pure chemical compounds were found to be a manner of manufacture even though the claim was, in essence, for a “wished for result” of a substantially pure compound falling within a claimed formula which could be created using various methods.
His Honour found that the claimed antibodies are not abstract information or mere desideratum, but compounds defined by their binding characteristics and ability to inhibit interactions between PCSK9 and LDLR. Importantly, his Honour noted that they were a product of human action (namely, the immunisation of genetically modified mice to generate an immune response, antibody isolation, and the subsequent creation of hybridomas for large-scale antibody production).
“Chilling effect” – Myriad
Sanofi argued that the invention’s subject matter did not fall within the established patentable categories, making the factors set out by the High Court in D’Arcy v Myriad Genetics Inc (2015) 258 CLR 334 relevant – in particular, the existence of any “chilling effect”. Sanofi argued that companies might discover only after extensive research into antibodies inhibiting PCSK9 binding to LDLR that their discovery was already patented.
His Honour held that Amgen’s invention fell within established patentable categories, so the factors in Myriad were not relevant. He also noted that any company making antibodies as described could rely on the broad exemption for experimental research under section 119C of the Act.
Fair basis
His Honour’s view on fair basis turned largely on whether the specifications described the same invention as the consistory clauses.
The consistory clauses described the invention as a class of antibodies that bind to epitopes comprising certain amino acid residues which either block or reduce binding of PCSK9 to LDLR. In contrast, the balance of the specification described the residues on the antibodies that, when interacted with or blocked, inhibited interaction between PCSK9 and LDLR. That is, the specification described the antibodies’ function by reference to amino acid residues, while the consistory clauses described that same function but by reference to epitopes.
Sanofi argued that this disconnect meant that fair basis requirement had not been satisfied.
His Honour held that it was enough that the specification provided scientific proof of the relationship between the amino acid residues and the PCSK9 epitopes. Justice Nicholas also rejected Sanofi’s complaint that the claims did not identify a specific antibody by its amino acid sequence; it was clear from the specification that the examples given were not exhaustive, and there was no need for Amgen to provide a definitive list.
Priority Date
Sanofi argued that the relevant priority documents lacked a specific description or disclosure of the relevant epitope or residues.
His Honour held that the specification followed “the same line of thought” as the priority documents, describing the invention without introducing new ideas or inventive steps. Justice Nicholas found the same regarding the Competition Claims, as the relevant priority document described an antigen binding protein that competes for binding with PCSK9 and included a description of the experiments used to assess competition.
Common General Knowledge
His Honour found that, at the priority date, it was common general knowledge (CGK) that:
- the main site at which PCSK9 degraded LDLR had not been established;
- it was not known whether PCSK9 acted on the outside of LDLR (i.e. extracellularly) or on its internal structure (i.e. intracellularly), and it was possible that PCSK9 used both mechanisms (this was important because antibodies only operate outside of cells); and
- the procedures and technologies available for making and screening antibodies were “reasonably mature” and would have been well known and routinely used by a person skilled in the art (PSA).
His Honour also found that, in relation to the inventive step analysis, two scientific papers (Lagace 2006 and Zhang 2007) also formed part of the CGK, a finding that is often difficult to establish.
Inventive Step
His Honour accepted that the PSA would have considered the possibility of PCSK9 inhibiting antibodies for treatment of hypercholesterolaemia and related cardiovascular disease. However, they would not have been naturally led to antibodies in the expectation of blocking or inhibiting binding between PCSK9 and LDLR, owing to the uncertainty in the CGK about whether PCSK9 acted extracellularly or intracellularly on LDLR. Specifically, at the priority:
- studies had established that PCSK9 reduced LDLR levels in genetically modified mice, suggesting that PCSK9-inhibiting antibodies could be developed, but further research was required to confirm whether and the extent to which PCSK9 operated extracellularly; and
- studies suggested an “emerging hypothesis” that PCSK9 functioned extracellularly, but the dominant view was that it was more likely that PCSK9 acted intracellularly on LDLR (which remained an open question even after the priority date).
Notably, his Honour rejected the evidence of Sanofi’s expert opinion that developing a PCSK9 inhibiting antibody would be straightforward or quick, as the expert’s opinion was “heavily influenced by hindsight” and the result of oversimplification. On this point, his Honour preferred the evidence of the expert that acknowledged and accounted for the impact of hindsight on what he believed to be known about PCSK9’s mechanisms at the time.
Lack of clarity
Sanofi argued that the Competition Claims were unclear because they did not define a numerical threshold for the word “competes” and failed to provide a workable standard to determine whether an antibody competes for binding.
Justice Nicholas rejected that argument in light of expert evidence that competition is a binary concept (an antibody either competes or it doesn’t) and that testing to identify competition between antibodies for binding on a protein such as PCSK9 would have been well known by the PSA based on the specification.
Sufficiency
Sanofi argued the invention had not been fully described because the claims were to individual antibodies. It relied on Tramanco Pty Ltd v BPW Transpec Pty Ltd (2014) 105 IPR 18, a previous decision of Justice Nicholas, which concerned a claim for a method of logging the performance of a vehicle suspension system. The method enabled three performance parameters to be determined. The Court said (in obiter) that the patentee rightly admitted the specification lacked sufficiency, as it didn’t enable the PSA to determine all three parameters using the method.
His Honour distinguished Tramanco because the claims in that case were to a method of producing alternative results or outcomes. By contrast, Amgen’s claims were not to distinct alternative outcomes, but to only one outcome (in the Residue and Epitope Claims, to the binding of antibodies and blocking of PCSK9, and in the Competition Claims, to competing antibodies).
His Honour also rejected the Sanofi’s argument that the claims were not for different inventions, but to a class of antibodies with certain functions (blocking, binding or competing).
Conclusion
Notably, to date, Australia is the only jurisdiction in which Amgen’s patents have been granted. With respect to the fair basis and sufficiency grounds, this may be attributed to the operation of pre-RTB Act principles. If that is right, this case provides a good working example of the effect of the RTB Act. It remains to be seen whether Sanofi will appeal this decision to the Full Court.